What do you think?
Rate this book
The myth of good AI: A manifesto for critical Artificial Intelligence
184 pages, Paperback
Published May 27, 2025
25 people want to read
About the author
Arshin Adib-Moghaddam
14 books11 followersArshin Adib-Moghaddam is Professor in Global Thought and Comparative Philosophies at SOAS, University of London and Fellow of Hughes Hall, University of Cambridge.
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 of 1 review
July 25, 2025
Here’s the (abbreviated) AI sermon I give to anyone who’ll listen (and some captive audiences, sure):
- LLMs are trained on racially and gender biased data. So, say you’re a “but what about AI and health tech?” cheerleader?
Well, that data is all based on data farmed from WASPs (mostly blokes), so your result will vary if you’re not melanin deprived. You can take a wild guess about how AI driven surveillance systems view non-white people, too.
- Training LLMs involves exploitation of poorly paid workers in the majority world, exposing them to the machine’s worst impulses and impacting their mental and physical health.
- Every AI startup admits there would be zero economic argument for its existence if it wasn’t for the fact they have been given carte blanche to carry off global IP heists to feed their models. They laugh at your “copyright”.
- The devastating mental health outcomes of chronic chatbot use are only starting to become apparent.
- Studies suggest that chatbot use by under eighteens is especially fraught.
- Reliance on the technology leads to a rapid devolution in critical and creative thinking, and catastrophically impacts basic problem solving and resilience.
- Even a simple ChatGPT query will use half a litre of water in the process of generating the answer to “who would win Buffy or Blade lol??”.
Globally, it is projected that AI use will consume the equivalent of Denmark’s annual water consumption by 2027, and generate as much carbon as New Zealand does annually.
- The data centres required to crunch your queries - which will number over 6000 globally by the end of 2025 - are built in minority communities and have quickly had devastating health impacts on them.
Tainted water supplies, shocking air pollution - these are all part of the Faustian bargain that is allegiance to the AI bullshit machine.
- Always remember: a large language model is just an incredibly complex spreadsheet of probable responses to your query, tweaked by human click workers based on the developer’s whims. We’ve been building variations on this theme since computers were the size of your spare room in the 1950s.
Artificial Intelligence is a marketing ruse, a consolidation of capital’s power over the worker post COVID, an authoritarian system of control and a distraction from the very real issues which should be uniting us.
No-one wants it, no-one truly needs it we should reject it at all costs.
Happy Saturday.
⚡️
- LLMs are trained on racially and gender biased data. So, say you’re a “but what about AI and health tech?” cheerleader?
Well, that data is all based on data farmed from WASPs (mostly blokes), so your result will vary if you’re not melanin deprived. You can take a wild guess about how AI driven surveillance systems view non-white people, too.
- Training LLMs involves exploitation of poorly paid workers in the majority world, exposing them to the machine’s worst impulses and impacting their mental and physical health.
- Every AI startup admits there would be zero economic argument for its existence if it wasn’t for the fact they have been given carte blanche to carry off global IP heists to feed their models. They laugh at your “copyright”.
- The devastating mental health outcomes of chronic chatbot use are only starting to become apparent.
- Studies suggest that chatbot use by under eighteens is especially fraught.
- Reliance on the technology leads to a rapid devolution in critical and creative thinking, and catastrophically impacts basic problem solving and resilience.
- Even a simple ChatGPT query will use half a litre of water in the process of generating the answer to “who would win Buffy or Blade lol??”.
Globally, it is projected that AI use will consume the equivalent of Denmark’s annual water consumption by 2027, and generate as much carbon as New Zealand does annually.
- The data centres required to crunch your queries - which will number over 6000 globally by the end of 2025 - are built in minority communities and have quickly had devastating health impacts on them.
Tainted water supplies, shocking air pollution - these are all part of the Faustian bargain that is allegiance to the AI bullshit machine.
- Always remember: a large language model is just an incredibly complex spreadsheet of probable responses to your query, tweaked by human click workers based on the developer’s whims. We’ve been building variations on this theme since computers were the size of your spare room in the 1950s.
Artificial Intelligence is a marketing ruse, a consolidation of capital’s power over the worker post COVID, an authoritarian system of control and a distraction from the very real issues which should be uniting us.
No-one wants it, no-one truly needs it we should reject it at all costs.
Happy Saturday.
⚡️
Displaying 1 of 1 review